You are on page 1of 3

Research letter

European Journal of Preventive


Cardiology

Primum non nocere: An updated 0(00) 1–3


! The European Society of
Cardiology 2019
meta-analysis on aspirin use in primary Article reuse guidelines:
sagepub.com/journals-permissions
prevention of cardiovascular disease DOI: 10.1177/2047487319826439
journals.sagepub.com/home/ejpc
in patients with diabetes

Federico Fortuni1, Gabriele Crimi2, Valeria Gritti2,


Alessandro Mandurino Mirizzi1, Sergio Leonardi1
and Gaetano M De Ferrari1

Individuals with diabetes have a two- to four-fold and as weighted sum of fatal equivalent of the above
increased risk of vascular events.1 A pivotal role in events using weights and methods previously reported.7
this setting is played by an exaggerated platelet activa- Random-effects risk ratios (RRs) were estimated using
tion. This has led to increasing interest over recent dec- a Mantel–Haenszel method with a person-year
ades in developing interventions aimed at reducing approach. Heterogeneity was calculated using the I2
cardiovascular risk in this population. Although the test and publication bias was visually assessed with
effectiveness of aspirin in secondary prevention of funnel plots. Univariate meta-regression for unadjusted
cardiovascular disease (CVD) in diabetes is established, log-RR was performed to explore the potential mod-
its use in primary prevention is still controversial2 and erator effect of mean age, aspirin dose, smoking status,
current European Society of Cardiology guidelines on hypertension, gender, type of diabetes, concomitant
CVD prevention do not provide a clear recommenda- statin therapy and compliance to study treatment on
tion.3 A number of randomized trials have reported on MACE. Statistical analyses were conducted using
the role of aspirin in primary prevention of CVD in Review Manager version 5.3 and OpenMeta-Analyst.
people with diabetes, but the majority of these studies Five trials met our inclusion criteria and were
were poorly powered with regard to the number of included in the analyses (N ¼ 24,037; mean follow-up
people with diabetes and reported results from 5.43  1.56 years; Table 1). Aspirin reduced the risk
subgroups.4 The recently published ASCEND trial pro- of MACE compared with placebo (1.5% vs. 1.7%
vided new evidences specifically assessing this issue in a person-year; RR: 0.90; 95% confidence interval (CI):
large population of diabetic patients.5 0.83–0.97; p ¼ 0.004; number needed to treat (NNT)
We performed an updated study-level meta-analysis over the mean follow-up ¼ 88). As shown in Figure 1
to evaluate the efficacy and safety of aspirin in primary the risk of MI, stroke, and cardiovascular and all-cause
prevention of CVD in diabetic patients. Electronic mortality did not differ between the two groups. Aspirin
databases were searched for clinical trials that rando- significantly increased the risk of major bleedings com-
mized diabetic patients without previous cardiovascular pared with placebo (0.53% vs. 0.43% person-year; RR:
events to aspirin or placebo. The process was per- 1.24; 95% CI: 1.06–1.45; p ¼ 0.03; number needed to
formed according to the PRISMA statement.6 Two
authors extracted data on patient characteristics and 1
Coronary Care Unit and Laboratory of Clinical and Experimental
outcomes. Efficacy outcomes were major adverse cardi- Cardiology - Fondazione IRCCS Policlinico San Matteo, Pavia, Italy
2
ovascular event (MACE), myocardial infarction (MI), Division of Cardiology - Fondazione IRCCS Policlinico San Matteo, Pavia,
stroke, and cardiovascular and all-cause mortality. Italy
Safety endpoints were major bleedings and malignan-
Corresponding author:
cies. Net adverse clinical events (NACEs) were com- Gabriele Crimi, Division of Cardiology, Fondazione IRCCS Policlinico San
puted as the unweighted sum of haemorrhagic stroke, Matteo, Piazzale Golgi 1, 27100 Pavia, Italy.
major extracranial bleeding, ischaemic stroke and MI Email: gabrielecrimi@gmail.com
2 European Journal of Preventive Cardiology 0(00)

Table 1. Study characteristics.

Patients Mean age Mean FU Male BMI Hypertension Statin use Aspirin Compliance
Study n (years) (years) (%) (kg/m2) (%) (%) dose (mg) (%)

ASCEND 15,480 63.3 7.4 62.6 30.7 61.6 75.3 100 70


ETDRS 3711 NA 5 56 NA 74.1 0 650 72.2
JPAD 2539 65 4.37 55 24 58 26 100 or 81 90
POPADAD 1276 60.3 6.7 44.1 29.25 NA NA 100 50
PPP 1031 64.2 3.7 48.2 29 60.5 12.7 100 71.8
ASCEND: A Study of Cardiovascular Events iN Diabetes; BMI: body mass index; ETDRS: Early Treatment Diabetic Retinopathy Study; FU: follow-up;
JAPD: Japanese Primary Prevention of Atherosclerosis with Aspirin for Diabetes; NA: not available; POPADAD: Prevention Of Progression of Arterial
Disease And Diabetes; PPP: Primary Prevention Project

Endpoints - n. of studies considered Estimate (95% CI) Events/ patient -years aspirin Events/patient -Years placebo

MACE - 5 0.898 (0.831, 0.970) 1201/78266 1338/78300


MI - 5 0.923 (0.822, 1.038) 539/78266 584/78300
Stroke - 5 0.923 (0.801, 1.063) 368/78266 399/78300
CV mortality - 5 0.909 (0.805, 1.026) 495/78266 545/78300
All-cause mortality - 5 0.943 (0.873, 1.018) 1241/78266 1317/78300
Major bleeding - 3 1.239 (1.061, 1.447) 354/67066 286/67131
Weighted NACE - 3 1.039 (0.946, 1.142) 866/67066 834/67131
Unweighted NACE - 3 1.032 (0.939, 1.133) 872/67066 846/67131
Cancer - 3 0.976 (0.849, 1.066) 960/63471 983/63445

0.8 0.97
1.45
Relative risk (log scale)

Figure 1. Efficacy and safety outcomes: aspirin versus placebo in primary prevention.
CI: confidence interval; CV: cardiovascular; MACE: major adverse cardiovascular event; MI: myocardial infarction; NACE: net adverse
clinical event

harm (NNH) over the mean follow-up ¼ 141). The risk updated meta-analysis suggests that aspirin does not
of malignancies and of NACE did not differ between the provide a clear and significant net clinical benefit
two groups. No publication bias was detected, both for when indistinctly used in primary prevention of CVD
efficacy and safety outcomes. These results were con- in patients with diabetes.
firmed also after excluding the Early Treatment
Diabetic Retinopathy Study, which included 26% of
Author contribution
patients with a previous cardiovascular event.8 The uni-
GC is responsible for the study rationale, manuscript
variate meta-regression did not show any moderator
drafting and final review. VG and AMM performed the lit-
effect on MACE of the variables considered.
erature search. FF and VG assessed the selected articles and
We must acknowledge some limitations. The defini- collected the data of interest. FF drafted the manuscript and
tion of major bleedings was not consistent across the performed the statistical analyses. All authors were involved
trials; however, we carefully revised and included in conception and the design of the study and interpretation
specific major haemorrhagic events in the NACE of the data, revised the manuscript critically for important
computation. The absence of any moderator effect in intellectual content and finally approved the manuscript
the meta-regression may be due to a type two error submitted.
related to the limited number of studies included in
the analyses. Declaration of conflicting interests
Aspirin significantly reduced by 10% the risk of
The author(s) declared no potential conflicts of interest with
MACE compared with placebo, whereas the reduction respect to the research, authorship, and/or publication of this
in the risk of individual hard efficacy endpoints did not article.
reach statistical significance. This modest benefit was
accompanied by a 24% increased risk of major bleed-
ings and no net clinical benefit was detected. Although Funding
the NNT to reduce a MACE in diabetic patients is The author(s) received no financial support for the research,
inferior to the NNH to induce a major bleeding, this authorship, and/or publication of this article.
Fortuni et al. 3

References diabetes: Meta-analysis of randomised controlled trials.


1. Emerging Risk Factors Collaboration. Diabetes mellitus, BMJ 2009; 339: b4531.
fasting blood glucose concentration, and risk of vascular 5. The ASCEND Study Collaborative Group. Effects of
disease: A collaborative meta-analysis of 102 prospective aspirin for primary prevention in persons with diabetes
studies. Lancet 2010; 375: 2215–2222. mellitus. N Engl J Med 2018; 379: 1529–1539.
2. Ridker PM. Should aspirin be used for primary prevention 6. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA
in the post-statin era? N Engl J Med 2018; 379: 1572–1574. statement for reporting systematic reviews and meta-ana-
3. Piepoli MF, Hoes AW, Agewall S, et al. 2016 European lysis of studies that evaluate health care interventions:
Guidelines on cardiovascular disease prevention in clinical Explanation and elaboration. Ann Intern Med 2009; 151:
practice: The Sixth Joint Task Force of the European w1–w30.
Society of Cardiology and Other Societies on 7. Renda G, di Nicola M and De Caterina R. Net clinical
Cardiovascular Disease Prevention in Clinical Practice. benefit of non-vitamin K Antagonist oral anticoagulants
Developed with the special contribution of the European versus warfarin in phase III atrial fibrillation trials. Am J
Association for Cardiovascular Prevention & Med 2015; 128: 1007–14.e2.
Rehabilitation (EACPR). Eur Heart J 2016; 37: 8. ETDRS Investigators. Aspirin effects on mortality and
2315–2381. morbidity in patients with diabetes mellitus. Early
4. De Berardis G, Sacco M, Strippoli GF, et al. Aspirin for Treatment Diabetic Retinopathy Study report 14. JAMA
primary prevention of cardiovascular events in people with 1992; 268: 1292–1300.

You might also like